A Troublemaker with Contagious Jailbreak Makes Chaos in Honest Towns

TL;DR

This paper introduces TMCHT, a large-scale multi-agent attack framework, and proposes ARCJ to effectively induce jailbreaks in multi-agent systems, highlighting security vulnerabilities in such environments.

Contribution

The paper presents TMCHT for evaluating multi-agent jailbreak attacks and introduces ARCJ, a novel method to enhance attack effectiveness in complex multi-agent topologies.

Findings

ARCJ improves attack success rates significantly.

TMCHT reveals vulnerabilities in multi-agent memory systems.

Proposes solutions for large-scale multi-agent security challenges.

Abstract

With the development of large language models, they are widely used as agents in various fields. A key component of agents is memory, which stores vital information but is susceptible to jailbreak attacks. Existing research mainly focuses on single-agent attacks and shared memory attacks. However, real-world scenarios often involve independent memory. In this paper, we propose the Troublemaker Makes Chaos in Honest Town (TMCHT) task, a large-scale, multi-agent, multi-topology text-based attack evaluation framework. TMCHT involves one attacker agent attempting to mislead an entire society of agents. We identify two major challenges in multi-agent attacks: (1) Non-complete graph structure, (2) Large-scale systems. We attribute these challenges to a phenomenon we term toxicity disappearing. To address these issues, we propose an Adversarial Replication Contagious Jailbreak (ARCJ) method, which optimizes the retrieval suffix to make poisoned samples more easily retrieved and optimizes the replication suffix to make poisoned samples have contagious ability. We demonstrate the superiority of our approach in TMCHT, with 23.51%, 18.95%, and 52.93% improvements in line topology, star topology, and 100-agent settings. Encourage community attention to the security of multi-agent systems.